Resource List: Shaping the Combatting Bias project¶
DOI: https://doi.org/10.5281/zenodo.16919549
This post has been updated from its first version (published November 2024).
This List of Resources1, as it is in its current form (August 2025), is the result of our initial enquiries into works, scholarship and practices around broad topics central to the Combatting Bias project: Bias, Dataset Creation and Documentation and the FAIR Principles.
Our aim: A bottom-up overview of concerns surrounding bias and related topics, as well as an insight into the scatteredness of these relevant resources. Collecting these resources into an accessible and shareable list — ideally editable as well, so it can reflect current interests and needs of those working with/on bias and so become community-driven. Furthermore, as knowledge does not merely exist in the realm of (peer-reviewed) academic papers and/or publications, we did our best to include films, exhibitions, fiction, and more. In this way we wanted this list to reflect a wide range of perspectives and approaches.
The outcome: An accessible List of Resources that reflects our initial research process. We were not able to find a good way to make the list openly editable, but we are grateful to those who reached out to us to recommend resources to add. Through this exercise, we’ve found that ‘bias’ can represent many interconnected themes and is often used as a heuristic.
This list, together with the interviews with our network, forms the theoretical basis of the Bias-Aware Framework, as well as provides concepts to include in the Bias Vocabulary. It also includes practical tools included in the Bias-Aware Guidelines.
We’re grateful to anyone who has contributed to this list, either directly or indirectly — these resources are truly the result of a collaborative effort. We hope that this list can be of use to those navigating and grappling with this concept of bias.
The resource list is available as (Excel) spreadsheet or Zotero library. Please note that there may be slight differences between the two as they have been created side-by-side (and not updated simultaneously).
The Challenge of “Sorting Things Out”¶
Each resource has a category(-ies), field(s), and concept(s) assigned to it.
Categories: The ‘categories’ are one of four broad themes, presented as subfolders (Bias, Dataset Creation and Documentation, FAIR, and Tools).
Fields: This refers to the (academic) field the resource is situated in. These, too, are broad categories and are also not mutually exclusive from one another. In the Zotero list, they are presented as tags with capitals.
Archaeology
Archival Sciences
Cultural Studies
Digital Humanities
Environmental and Life Sciences
Fiction
Heritage Studies
History
Law
Literary Studies
Machine Learning
Natural Language Processing
Open Science
Philosophy
Science and Technology Studies
Social Sciences
Sociology
Concepts: General and specific concepts encountered in the resources. These are presented through tags using all lower-case.
absences
accessibility
accountability
against the grain
AI
anti-racism
archival turn
art
assessment
care
CARE
caste
categories of bias
categorisation
category of analysis
citizen science
class
classification
collaboration
collections as data
community engagement
community memory
community relationship
complexity
conscientiousness
contentiousness
contextualisation
control
critical fabulation
cross-cultural selection criteria
crowdsourcing
cultural colonialism
cultural memory
curation
data
data justice
data sovereignty
decoloniality
decolonisation
deconstruction
deepfake
description length
digital legacy
digitisation
dignity
discrimination
disinformation
documentation
enchantment
ethical description
ethics
FAIR
fairness
gender
generative AI
handwritten text recognition
harmful language
heteronormativity
historicity
honesty
identity
inclusive description
inclusivity
indigeneity
information politics
interoperability
interpretation
intersectionality
justice
knowledge
knowledge as a commons
knowledge production
language
layout analysis
legacy description
memory
metadata
noisiness
objectivity
offensive language
open science
othering
ownership
paradata
performance
plurivocality
polyvocality
positionality
power
privacy
race
radical empathy
reflexivity
regulation
reparation
representation
representativeness
reproducibility
research process
responsibility
restitution
rewards
sexuality
silence
silences
slavery
social justice
standard
statistics
student activism
technology of recovery
third spaces
transcription
transdisciplinary
transparency
trust
uncertainty
usefulness
user needs
visualisation
From how overwhelming this list of concepts is, it is clear that the concept-tags do not necessarily categorise the resources in a workable way. The goal was also not to achieve a formal framework for categorisation for these resources; instead, it meant to show the whole range of concepts that can be associated with ‘bias’ (also through FAIR and documentation).
This range of concepts shows that bias is a ‘heuristic’, which can be a productive category of analysis in research processes. Please note that not all concepts outlined above are incorporated into the Bias Vocabulary, as this list represents our initial orientation for our research.
Credit: Stadsarchief Amsterdam / Internationaal Persfoto Bureau N.V.
Methodology¶
We did not have a systematic methodology for the compilation of resources. The selection-criteria for including resources to the list was their relevance to one of the (initially) three categories: Bias, FAIR and Dataset Creation & Documentation. This relevance could either be direct (e.g. a philosophical paper outlining different ‘varieties of bias’: Johnson, 2024) or indirect (e.g. an artist’s video aiming to traverse the gap existing in the archives on the lived experiences of children sent from the Danish West Indies to Denmark and displayed at the colonial exhibition: La Vaughn Belle, 2022).
The category for ‘Tools’ was added while we matched the list from the Google Sheets to the current Zotero list. This category is intended to distinguish theoretical resources from practical ones, aimed to be useful to researchers looking specifically for resources that can support them when engaging, explaining and/or visualising biases.
We adhered to the following descriptions of each category, adding resources to one or multiple categories:
| Category | Description |
| Bias | Resources that engage with the concept of ‘bias’, in all its forms, either direct or indirectly. |
| FAIR | Resources that use, explain, and/or build on the FAIR principles. These will focus on data - and how it is being produced, published, and stored. |
| Dataset Creation & Documentation | Resources that discuss the process of creating a dataset (including challenges and unanswered questions). |
| Tools | Resources that are not theoretical and/or conceptual in nature, but provide concrete and practical workflows, guidelines, or templates that can help researchers engage with bias (e.g. checklists, datasheets). |
The first resources were added based on ‘expertise knowledge’ on data ethics and search queries of ‘bias in [discipline, e.g. LLMs, digital humanities]’. From these, we snowballed into new resources, through bibliographies and references of already added resources and recommendations from colleagues, social media, events and email.
The concepts were mostly collected through descriptive annotation (in contrast to in vivo annotation = literally cited) - assigning a relevant phrase, known to us through interviews or reading different literature (e.g. ‘harmful language’ or ‘plurivocality’) to describe the content of the resource, even if this phrase was not literally used. This was so to balance bottom-up research with at least some form of systematic consistency.
Because of the lack of systematic methodology for collating materials, the list we created is far from comprehensive. Our interests and networks are represented in the list. Our choice to make it public was therefore based on this awareness, inviting more perspectives to engage with these topics and to add to our resources. The list in its current form is a transparent insight into the development of our thinking and basis of the Bias-Aware Framework. Please do not use or interpret this list as a comprehensive overview.
-
We’ve been inspired by Casey Fiesler’s spreadsheet on AI Ethics and Policy news ↩
-
Where possible, we’ve included links to open access publications that are not behind paywalls. However, we haven’t always been able to secure these for every resource. ↩